search

charlesq34 / pointnet

PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation
Other
4.82k stars 1.46k forks source link

Loss in advanced model #60

Closed FredHaa closed 6 years ago

FredHaa commented 6 years ago

Hi,

I'm using PointNet for semantic segmentation, and would like to use the advanced model instead of the basic.

tf.concat has changed parameter order, so you need to change concat_feat = tf.concat(3, [point_feat, global_feat_expand]) on line 64 to concat_feat = tf.concat(axis=3, values=[point_feat, global_feat_expand]) for the model to work.

However, I cannot figure out how to use the loss as it takes 3 arguments instead of the basic models 2.

So my question is, what am I supposed to pass as the end_points argument?

FredHaa commented 6 years ago

Disregard this, the end points are returned by the model